I want to preserve publicly submitted photos, videos, and firsthand accounts from a series of largely peaceful student protests where people have reported beatings, tear gas, excessive force, sexual assault, and difficulty identifying some security personnel. The goal is a durable, organized archive that could assist journalists, lawyers, human-rights organizations, or courts—not harassment, doxxing, or vigilantism.
I'm considering secure uploads, anonymous submissions, optional metadata preservation, encrypted storage, duplicate detection, searchable timelines and maps, audit logs, review labels such as verified or disputed, and controlled exports for authorized organizations. I have very little web-development experience and would like to know where to begin.
What language and framework would be sensible to learn? Can AI coding tools help without creating serious security risks? What are the main threats, how should large videos and backups be stored, how can contributors be protected, and how can uploaded files be checked for tampering? Are there existing open-source evidence-preservation projects worth studying?
2 Answers
The programming language matters less than the architecture and operational discipline. Choose a well-supported stack you can learn—Node.js, C#, Python, or another mainstream option—and first build a harmless prototype with fake data. Learn HTTP, databases, authentication, authorization, secure file handling, backups, and basic threat modeling before accepting real submissions.
For file integrity, calculate a cryptographic hash when each file is received and store the hash and submission record in a separately protected, append-only audit system. Keep several encrypted copies in independently secured locations, with access keys separated from the storage. A hash can show that a stored file changed, but it cannot prove that the original recording was authentic or that its metadata was never altered before upload; chain of custody and documented review are equally important.
Do not automatically publish locations, faces, names, or raw metadata. Use a trained review team, minimize collection, establish retention and deletion policies, and consult local lawyers and established digital-evidence or human-rights organizations. Existing whistleblower and evidence-preservation tools can provide ideas, but adapting them safely still requires a professional security assessment.
For a project handling sensitive evidence and potentially identifying witnesses, start with a threat model and professional guidance before writing code. The biggest risks may be compromised phones, exposed server logs, weak authentication, abusive access, metadata leaks, malware in uploads, and legal obligations around sexual-assault reports and personal data.
A practical first version should be much smaller: controlled intake, encrypted storage, strict access permissions, immutable audit records, regular offline backups, and a careful review process. Keep the upload service separate from the public-facing site, scan and transcode media in an isolated environment, and avoid collecting IP addresses or unnecessary device information. Anonymous submission is difficult to guarantee, especially if the uploader’s device or network is compromised, so do not promise perfect anonymity without expert review.
AI can help explain technologies, write tests, and produce prototypes, but it cannot reliably design or verify security for a high-stakes system. Treat generated code as untrusted until experienced developers and security professionals review it.
Storing everything permanently can also create danger for contributors. A clear consent process, restricted exports, redaction tools, and a way to correct or remove material when legally necessary should be designed from the beginning.